[erlang-questions] Can Mnesia replace conventional RDBMS?

Chris Hicks <>
Sun Jun 2 17:39:10 CEST 2013


Also, Mnesia focuses on the CA in CAP, and doesn't behave very nicely when
you get a split cluster that you then have to bring back into line. The
pain of changing the records can be mitigated somewhat by having a single
base record type that you store that's basically just a key/value pair
(-record(document, {doc_id, doc}) for example) with your changing data held
in the value part...but that makes a lot of operations just a bit more
complicated than they otherwise would be and you have to do a lot more work
yourself.

If you have a system design which isn't going to change and is going to
remain on a single server Mnesia might just work for you, but if you're
looking at any sort of distributed or changing application architecture you
might want to consider another solution.


On Sun, Jun 2, 2013 at 1:19 AM, Yash Ganthe <> wrote:

> The Mnesia documentation says:
> largest possible mnesia table (for now) is 4Gb
>
> A table can be distributed across nodes so that part of it remains on one
> node and part on another. In a distributed sense, does this mean 4GB per
> table per node, or 4GB per table across nodes?
>
> Is there a way by which Mnesia can be used in a system that will store
> about 100GB of data in some tables and support ACID properties just like
> popular RDBMS's like MySQL/Oracle?
>
> Thanks,
> Yash
>
> _______________________________________________
> erlang-questions mailing list
> 
> http://erlang.org/mailman/listinfo/erlang-questions
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20130602/cd8676a6/attachment.html>


More information about the erlang-questions mailing list